The long term goal of this project is to understand the underlying mechanisms that cause fatty liver (NAFLD) and nonalcoholic steatohepatitis (NASH) in obesity/type 2 diabetes, and how such mechanisms relate to systemic insulin resistance. We also seek to unravel the paradox that while NASH is tightly correlated with insulin resistance in obese humans and mice, these two syndromes are clearly dissociated in certain gene KO mouse models. The central hypothesis of this proposal solves this riddle by positing that hepatocyte cytosolic Acetyl CoA levels promote NAFLD and NASH through producing palmitate/cholesterol toxicity, while hepatocyte mitochondrial Acetyl CoA levels drive insulin resistance by stimulating pyruvate carboxylase and gluconeogenesis. Thus, we propose that while hepatocyte Acetyl CoA pools are often coordinately elevated, they can be disconnected in certain genetic mouse models of obesity and fatty liver. In order to test our hypothesis and attack this problem directly, we apply novel gene silencing technology that combines unique RNA modifications and GalNAC-directed hepatocyte targeting in ?self delivery? RNAi (sdRNA) compounds. These compounds can silence single or multiple targeted hepatocyte genes for 2 months or more after a single subcutaneous injection in mice. Using GalNAC-sdRNA, we can selectively target and silence each of the multiple pathways that produce cytosolic Acetyl CoA (e.g., ACLY and ACSS2 pathways) versus mitochondrial Acetyl CoA(e.g., FATP2/5 pathway), while avoiding prohibitive costs and time in generating multiple gene KO mice. In Aim 1, we couple this powerful RNAi technology with a novel method that quantifies hepatocyte mitochondrial Acetyl CoA vs total cellular Acetyl CoA to determine the relative contributions of ACLY, ACSS2 and FATP2, FATP5 to these specific hepatocyte Acetyl CoA pools in lean and HFD mice. In Aim 2 we propose to deplete hepatocyte cytosolic Acetyl CoA in NAFLD/NASH mouse models by appropriate GalNAC-sdRNA gene targeting learned from Aim 1, and determine its impact on liver triglyceride, inflammation, fibrosis and glucose tolerance as well as its impact on Kupffer and Stellate cell dysfunction. For example, we will test whether depletion of hepatocyte Acetyl CoA levels in NASH mouse models attenuates collagen production by Stellate cells through downregulation of hepatocyte transcription factor TAZ, which drives hepatocyte Indian hedgehog (IHH) secretion and Stellate activation. These studies will also resolve the key question whether Kupffer and Stellate cell dysfunction is driven by hepatocyte NAFLD versus independently promoted by circulating factors, or both. Finally, in Aim 3 we will test a potential therapeutic strategy by determining whether GalNAC-sdRNAs targeting multiple hepatocyte genes will simultaneously alleviate all three syndromes of NAFLD, NASH and insulin resistance in obesity/type 2 diabetes. This approach has major clinical advantages since multiple GalNAC-sdRNAs against different genes consist of the same chemical composition and, unlike small molecules, are evaluated for use in the clinic as a single therapeutic agent.